1. Field of the Invention
The invention disclosed in this specification relates to optical apparatuses utilizing a laser such as apparatuses for performing an annealing process by means of irradiation with laser beams (laser annealing apparatuses) and, more particularly, to an laser annealing apparatus projecting beams with a large area which is capable of providing an uniform effect of irradiation. Such a laser annealing apparatus is used in semiconductor manufacturing steps and the like.
Laser beams with a large area are used in apparatuses including exposure apparatuses for forming fine circuit patterns such as semiconductor circuits. Especially, ultraviolet laser beams are used for forming circuits with design rules on a sub-micron basis.
2. Description of the Related Art
A techniques for crystallizing amorphous silicon films by irradiating them with laser beams has been known. Another known laser irradiation technique is to irradiate silicon films with laser beams in order to recover them from damage to crystallinity thereof due to the implantation of impurity ions and in order to activate the implanted impurity ions. Such techniques are referred to as "laser annealing techniques".
A typical example of the latter technique is the annealing of the source and drain regions of a thin film transistor. Those regions are annealed by irradiating them with laser beams after ions of impurities, typically phosphorus or boron, are implanted into those regions.
Such a process of irradiation with laser beams is characterized by the fact that there is substantially no thermal damage to a substrate.
The feature of giving no thermal damage to a substrate reduces limitations on the materials to be subjected to such a process and is advantageous, for example, in forming a semiconductor device on a substrate made of glass or the like which has low heat resistance. This feature is especially important in the fabrication of active matrix liquid crystal displays which recently have an increasing range of application.
For an active matrix liquid crystal display, it is desirable to use a glass substrate from the viewpoint of cost and the requirement for a larger surface area.
A glass substrate can not withstand a heating process at temperatures as high as 600.degree. C. or more or 700.degree. C. or more. An effective technique for avoiding this problem is to perform annealing after the crystallization of a silicon film and the implantation of impurity ions as described above by irradiating it with laser beams.
Even when a glass substrate is used, a method employing irradiation with laser beams results in substantially no damage to the glass substrate. It is therefore possible to use a glass substrate in fabricating a thin film transistor having a crystalline silicon film.
There has been another proposal to use laser beams as a light source for forming fine circuit patterns taking advantage of the fact that laser beams are coherent light. Especially, the use of an ultrasonic laser makes it possible to obtain fine patterns having sizes in sub-microns or smaller.
However, since laser beams have small beam areas when they are generated by a laser apparatus (hereinafter they are referred to as "source beams"), it is common to process a large surface area by scanning laser beams across it. This results in problems including low uniformity of the effect of the process in a surface and a long period of time required for the process. Especially, common source beams result in a significant problem from the viewpoint of uniformity of the effect of processing when used as they are because they have non-uniform distribution of light intensity.
Under such circumstances, a technique has been proposed wherein source beams are processed into beams having highest possible uniformity and the beam size is changed in accordance with the shape of the surface area to be processed and the like. Common beam shape is rectangular or linear shape. Such an arrangement makes it possible to perform uniform laser annealing over a large surface area.
FIG. 1A shows an example of a laser irradiation apparatus in which source beams are processed. For example, an excimer laser is used as the laser oscillator. Laser beams are oscillated by decomposing predetermined gases by means of RF discharge to produce an excited state referred to as "excimer state".
For example, in a KrF excimer laser, an excited state KrF* is obtained by high voltage discharge using Kr and F as raw material gases. While this excited state is unstable as indicated by its duration in the range from several nano-seconds to several micro-seconds, KrF in the ground state is more unstable. This results in inverted population wherein the density in the excited state is higher than the density in the ground state. As a result, induced radiation occurs, which makes it possible to obtain laser beams having relatively high efficiency.
The laser oscillator is not limited to an excimer laser, and other pulse lasers or continuous lasers may be used. In general, pulse lasers are appropriate for the purpose of achieving a high energy density.
As shown in FIG. 1A, a source beam emitted by the laser oscillator is processed into an appropriate size by a beam expander formed by a concave lens or a convex lens.
The beam then enters an optical device referred to as "homogenizer" which includes at least one lens device (multi-cylindrical lens) having a multiplicity of cylindrical lenses (generally in a parabolic configuration). As shown in FIG. 1B, a conventional multi-cylindrical lens includes a plurality of cylindrical lenses 1 through 5 (which are all convex lenses) formed on a single sheet of glass.
In general, two multi-cylindrical lenses are used and arranged so that they are perpendicular to each other. Obviously, the number of the multi-cylindrical lens may be one or three or more. When one multi-cylindrical lens is used, the non-uniformity of a source beam in one direction is dispersed. When two or more multi-cylindrical lenses are formed in the same direction, the same effect as increasing the number of the cylindrical lenses can be achieved.
When a beam passes through the multi-cylindrical lens, the beam can be converted into a uniform beam having a distributed energy density. The principle behind this will be described later. Thereafter, the beam is processed by a converging lens into a desired shape or, if needed, deflected by a mirror to be projected upon a sample (see FIG. 1A).
A description will now be made on the principle of a conventional homogenizer (multi-cylindrical lens) and a problem of the same which is the problem to be solved by the invention. In order to avoid complication, discussion on an optical basis will be focused on only one surface. Laser beams that have passed through a multi-cylindrical lens are as shown in FIG. 2A.
Here, the multi-cylindrical lens L includes five convex cylindrical lenses 1, and the beam incident upon each of the cylindrical lenses is refracted by the cylindrical lens. After being converged at a focal point, the beams are diffused. This process results in a region in which all of the beams that pass through the respective cylindrical lenses are mixed (mixed region).
Let us assume here that the distribution of the optical intensity of the beams is polarized, resulting in differences in the intensity of the beams incident upon the respective cylindrical lenses. In the mixed region, however, such polarization is scattered because the beams that pass through the respective cylindrical lenses are mixed. That is, the optical intensity is made uniform. It is thus possible to obtain beams having less distribution of optical intensity (see FIG. 2A).
When we look at the paths of the beams that pass through the multi-cylindrical lens, the beams can be regarded as beams emitted from point light sources F (i.e., focal points) arranged at equal intervals (distances "a") (see FIG. 2B).
The same effect can be achieved by providing a convex cylindrical lens 1.sub.1 on one side of a glass substrate and a convex cylindrical lens 1.sub.2 on the other side at an interval "a". In FIG. 3A, the path of a beam that has passed through the cylindrical lens 1.sub.1 is indicated by the solid line, and the path of a beam that has passed through the cylindrical lens 1.sub.2 is indicated by the broken line. In this case, a mixed region is obtained as in the case shown in FIG. 2A (see FIG. 3A).
When we look at the paths of the beams that pass through the multi-cylindrical lens, as shown in FIG. 3B, the beams can be regarded as beams emitted from two kinds of point light sources F.sub.1 and F.sub.2 (i.e., focal points) (see FIG. 3B).
In the case of the conventional multi-cylindrical lens as described above, since there is an angle at the ends of each cylindrical lens (boundaries between itself and other cylindrical lenses), beams have been scattered at such regions, which has resulted in optical loss. This means that the laser beams can not be effectively utilized and the beam intensity is reduced.